| Abstract: | A method is presented for the analysis of the risk to man or other systems by effects caused by a combination of two insults. The approach is purely phenomenological and has no built-in restrictions other than those required by logic and those imposed by the limited set of trial functions. The analysis attempts to isolate the dominant contributions to the risks from either agent alone and from both agents in combination. To demonstrate the viability of the method, it is applied to the analysis of human health data on oral and esophageal cancer correlated with alcohol and tobacco consumption. For both cancers, it is found that the most probable form of the relative risk is separable into a factor related to tobacco and one related to alcohol. Both have the same algebraic form and the values of the risk coefficients are practically equal. An analysis of analogous but limited data on laryngeal cancer cannot assign an algebraic form of the risk function but still yields comparable risk coefficients. Thus it is possible that all three cancers have the same risk function as a function of alcohol and tobacco consumption. |
